Anda di halaman 1dari 12

# Utah State University

DigitalCommons@USU
Foundations of Wave Phenomena Library Digital Monographs
1-1-2012
15 Schrodinger Equation
Charles G. Torre
Department of Physics, Utah State University, Charles.Torre@usu.edu
Part of the Physics Commons
htp://digitalcommons.usu.edu/foundation_wave/8
Tis Book is brought to you for free and open access by the Library Digital
Monographs at DigitalCommons@USU. It has been accepted for inclusion
in Foundations of Wave Phenomena by an authorized administrator of
digitalcommons@usu.edu.
Recommended Citation
Torre, Charles G., "15 Schrodinger Equation" (2012). Foundations of Wave Phenomena. Book 8.
htp://digitalcommons.usu.edu/foundation_wave/8
15. The Schrodinger Equation.
An important feature of the wave equation is that its solutions q(r, t) are uniquely
specied once the initial values q(r, 0) and q(r, 0)/t are specied. As was mentioned
before, if we view the wave equation as describing a continuum limit of a network of
coupled oscillators, then this result is very reasonable since one must specify the initial
position and velocity of an oscillator to uniquely determine its motion. It is possible to
write down other equations of motion that exhibit wave phenomena but which only
require the initial values of the dynamical variable not its time derivative to specify
a solution. This is physically appropriate in a number of situations, the most signicant of
which is in quantum mechanics where the wave equation is called the Schrodinger equation.
This equation describes the time development of the observable attributes of a particle via
the wave function (or probability amplitude) . In quantum mechanics, the complete
specication of the initial conditions of the particles motion is embodied in the initial
value of . The price paid for this change in the allowed initial data while asking for
a linear wave equation is the introduction of complex numbers into the equation for the
wave. Indeed, the values taken by are complex numbers. In what follows we shall explore
some of the elementary features of the wave phenomena associated with the Schrodinger
equation.
15.1 One-Dimensional Schrodinger equation
Let us begin again in one spatial dimension, labeled by x. We consider a complex-
valued function . This is a function that associates a complex number (x, t) to each
point x of space and instant t of time. In other words, at each (x, t), (x, t), is a complex
number. Consequently, we can if desired break into its real and imaginary parts:
(x, t) = f(x, t) + ig(x, t), (15.1)
where f and g are real functions. We can also use a polar representation:
(x, t) = R(x, t)e
i(x,t)
, R 0. (15.2)
See 1.3 for a review of complex variable notation.
The complex-valued function is called the wave function youll see why shortly.
The wave function is required to satisfy the Schrodinger equation:

h
2
2m

x
2
+ V = ih

t
. (15.3)
Here V = V (x, t) is some given real-valued function of space and time representing the
potential energy function of the particle, h is Plancks constant (h) divided by 2, and
131
m is a parameter representing the mass of the particle. The Schrodinger equation is
a complex, linear, homogeneous, partial dierential equation with variable coecients
(thanks to V (x, t)). It is equivalent to a pair of real, coupled, linear dierential equations
for the real and imaginary parts of as you can see by using the fact that equality of
complex numbers means separate equality of their real and imaginary parts (exercise).
The Schrodinger equation species the time evolution of a quantum mechanical par-
ticle,* thus it plays a role in quantum mechanics roughly akin to the role played by the
famous
~
F = m~a in Newtonian mechanics. While we often speak of the Schrodinger equa-
tion, strictly speaking there is no single dierential equation valid for all situations. Rather,
each potential energy function denes a Schrodinger equation appropriate to the physical
system. This is also true with
~
F = m~a in Newtonian mechanics; one uses dierent forms
for
~
F depending upon the physical situation. We also note that in some applications of
the Schrodinger equation it is useful to allow the potential energy function V to be com-
plex valued. Such potentials can be used to model processes involving particle decay. For
simplicity we shall assume that the potential energy function is real. (See the Problems
and also 15.4 for hints as to what happens when we let V be complex-valued.)
While we certaintly wont be ofering a course in quantum mechanics in this text, it is
worth commenting on the physical meaning of solutions to (15.3). The simplest use of the
wave function is via the rule that

## (x, t)(x, t) dx is the probability that a measurement

of the particles position at time t will nd the particle in the region between x and x+dx.
More complicated expressions involving are used to give the probability distributions
for other particle observables. You will have a chance to get used to such ideas in a later
course in quantum mechanics. Fortunately, we do not really need to understand much of
quantum mechanics in order to see the basic wave phenomena embodied in the Schrodinger
equation. Still, from time to time it will be appropriate to make a few remarks concerning
the physical interpretation of some of our results.
15.2 Free Particle Solution of the Schr odinger Equation
Let us now try to understand the sense in which (15.3) is a wave equation. This is most
easily done by considering the special case V (x, t) = 0, which physically corresponds to
the motion of a free particle. Although you can probably guess solutions to this equation,
let us apply some of the techniques we have developed.
We begin with separation of variables; we try a solution of the form
(x, t) = X(x)T(t), (15.4)
* There is also a Schrodinger equation for systems of particles, not to mention even more
exotic dynamical systems. But we will stick to the simplest case of a single particle.
132
and substitute to nd (exercise):

h
2m
X
00
X
= i
T
0
T
. (15.5)
As usual, we conclude that

h
2m
X
00
= X, (15.6)
and
iT
0
= T,
where is some constant. These equations are easily solved:
X(x) = Ae
ikx
, T(t) = Be
i(k)t
, (15.7)
where k is any constant and
(k) =
hk
2
2m
. (15.8)
Note that we could have written X(x) = Ae
ikx
, but we can get both choices of sign by
choosing k positive or negative, so for simplicity we drop the . Keep in mind, though,
that for a given there are two independent solutions for X(x), namely e
i|k|x
.
Since is complex-valued, there is no obvious a priori restriction on whether k is
real or not. As it happens, physical considerations in conjunction with the principles of
quantum mechanics end up requiring k to be real in this example, so well only consider
that case. The solution
(x, t) = Ce
i(kx(k)t)
(15.9)
is then a complex form of a traveling wave (i.e., its real and imaginary parts are traveling
waves). We do not need to take the real part of , however, since the wave function
is allowed to be complex. Like the wave equation, the Schrodinger equation is linear and
homogeneous. This means that one can take linear combinations of solutions (with complex
coecients) to get new solutions a fact that has far-reaching physical consequences in
quantum mechanics. The general solution of the free particle (V (x, t) = 0) Schrodinger
equation is a familiar superposition of traveling waves:
(x, t) =
1

2
Z

C(k)e
i(kx(k)t)
dk. (15.10)
As a homework problem you will derive this same form of the solution using Fourier
transform methods.
Let us make a few comments regarding the physical meaning of (15.9) and (15.10).
Physically, the wave function (15.9) represents a free particle with momentum p = hk
and energy E = h =
h
2
k
2
2m
. Recall that

## (x)(x)dx is interpreted as the probability

133
for nding the particle between x and x + dx. This probability is the same throughout
all space for a particle described by (15.9) because

= constant, independent of x
(exercise). Thus the particle in a state described by (15.9) has an equal chance to be
found anywhere in space. By contrast, the momentum of the particle in this state is known
with certainty to have the value hk. This state of aairs is an extreme manifestation of the
position-momentum uncertainty principle: the statistical spread in the position is inversely
proportional to that of the momentum. Thus, in particular, if the momentum is known
precisely (vanishing statistical uncertainty) then the position takes all values with equal
probability. The general solution (15.10) of the free particle Schrodinger equation, being a
superposition over plane waves, corresponds to a superposition of momenta and energies.
Because of this superposition, neither the energy or momentum of a free particle described
by (15.10) has a precise value in the sense that there is a probability distribution describing
the range of possible outcomes of a measurement of these observables.
Exercises: What choice of C(k) in (15.10) corresponds to (15.9)? What is the wavefunction
at t = 0 if we set C(k) = 1?
Equation (15.8) denes the relation between frequency and wave number (equivalently,
wavelength) for solutions of the free particle Schrodinger equation. It is the dispersion
relation for this Schrodinger equation. Compare the dispersion relation for the Schrodinger
equation with the dispersion relation (8.67) for the wave equation in one dimension. The
latter exhibits a linear relation between frequency and wave number while the former
exhibits a quadratic relation. To understand the implications of these dierent dispersion
relations let us recall that, in general, sinusoidal waves of the form Asin(kx t) travel
with speed given by /k. For the wave equation, (8.67) tells us that /k = v, i.e., the
waves travel with speed v irrespective of the frequency (or wavelength) of the wave. For
the Schrodinger equation, (15.8) tells us that

k
=
hk
2m
, (15.11)
which implies that the speed of the sinusoidal wave depends upon the wavelength.* Thus
waves with dierent wavelengths travel at dierent speeds, the shorter wavelengths having
the higher speeds (exercise). At any given time we can Fourier analyze any solution of the
free particle Schrodinger equation into a superposition of sinusoidal waves with varying
wavelengths (see (15.10)). Since each of these waves travels with a dierent speed, the
shape of the wave will not be preserved in time as it is in the case of the wave equation
* Note that this result says the sinusoidal wave speed is one half the momentum hk of the
particle divided by the mass! So one cannot interpret the free particle wave motion as
particle motion; the slogan particles are waves has to be handled with care.
134
the Schrodinger wave will in fact disperse as the shorter wavelengths outrun the
longer wavelengths (see g. 21 below). This is the origin of the term dispersion relation
for formulas such as (8.67) and (15.8).
In contrast to the wave equation, the general solution to the Schrodinger equation
involves only one undetermined complex function of one variable. We see this explicitly
in the free particle case (15.10), where the undetermined function is represented by C(k).
This reects the fact that only the initial value of the wave function (x, 0) is needed to
uniquely x the solution. Thus suppose (x, 0) = f(x), where f(x) is some given function.
Then C(k) is the Fourier transform of f(x) (exercise). According to the rules of quantum
mechanics, C(k) denes the probability amplitude for momentum, that is, C

(k)C(k)dk
is the probability for nding momentum between hk and hk + hdk. Let us illustrate this
with an example which we have already explored mathematically.
Consider the initial condition
(x, 0) = Ae

x
2
a
2
. (15.12)
Physically, this corresponds to a particle which is most likely found at the origin, but has
a non-vanishing probability to be found anywhere on the x-axis. The likelihood for nding
the particle away from the origin grows as the parameter a is increased, i.e., as the width
of the Gaussian increases. From our previous work with the Gaussian prole, you can
check that its Fourier transform is (exercise)
C(k) =
1
2

2
Aae

k
2
a
2
4
. (15.13)
C(k) (and hence |C(k)|
2
) is also a Gaussian. Evidently, the momentum is most likely to
be zero in this state, but the likelihood for nding a non-zero momentum increases as the
parameter a decreases. The probability distribution in position has its width increasing
with increasing a, while the probability distribution in momentum, has its width decreas-
ing with increasing a. This is a good example of the uncertainty principle for position
and momentum: as the probability distribution in position (momentum) becomes more
tightly localized around a given value the probability distribution in momentum (position)
becomes more de-localized. Speaking more loosely, as the position of the particle becomes
more (less) uncertain the momentum of the particle becomes less (more) uncertain.
135
Figure 21. Time dependence of the Gaussian wave-packet solution to the
Schr dinger equation. In each graph ( ) ! Re is the dashed line, ( ) ! Im is the
dotted line, and ! ! * is the solid line. Note that the time dependence is
different than for the wave equation.
10 0 10
0.5
0
0.5
1
TIME = 0
10 0 10
0.5
0
0.5
1
TIME = 1.5
10 0 10
0.5
0
0.5
1
TIME = 5
x
136
15.3 The 3-Dimensional Schrodinger Equation
The generalization of the Schrodinger equation for a particle to three dimensions in-
volves the Laplacian again:

h
2
2m

2
+ V (r, t) = ih

t
. (15.14)
Now, of course, the wave function depends on the position in three-dimensional space,
r = xi +yj +zk and the time t, = (r, t). You can easily see that this equation reduces
to (15.3) if the y and z dependence of and V are eliminated. The meaning of the wave
function is a simple extension of the 1-d result: |(r, t)|
2
d
3
x is the probability that the
particle is found in a volume element d
3
x at the point r at time t. The free particle case
(V = 0) is easily treated by separation of variables and/or Fourier methods, as you will
explore in a homework problem.
15.4 Conservation of Probability, Normalization
The Schrodinger equation admits a very important conservation law, which provides
a nice example of the continuity equation formalism we discussed earlier. To derive the
conservation law, we need both the Schrodinger equation (15.14) and its complex conjugate

h
2
2m

+ V (r, t)

= ih

t
. (15.15)
(Note: Here we have used the assumption that the potential energy is a real function.) We
can construct a continuity equation as follows. Multiply the Schrodinger equation (15.14)
by

and multiply the complex conjugate equation (15.15) by . Take the dierence of
the two resulting equations to get (exercise)
ih

t
+

h
2
2m

2

2

= 0. (15.16)
Thus, if satises (15.14) (with V real), then it also satises (15.16). Next, we recall the
identity (10.3). Let us apply (10.3) to the vector eld

:
(

) =

+
2

. (15.17)
Similarly
(

) =

2
. (15.18)
Subtracting these two results and using the fact that the dot product is commutative
(A B = B A) we get (exercise)
[

] =
2

2
. (15.19)
137
Thus the second term in parenthesis in (15.16), involving the Laplacian, can be expressed
as the divergence of a vector eld. It is straightforward to check that the rst term in
parenthesis in (15.16), involving the time derivative, can be expressed as a time derivative
(exercise):

t
+

t
=

t
(

). (15.20)
From these manipulations we see that the result (15.16) can be expressed as a continuity
equation (exercise)

t
+ j = 0 (15.21)
where
=

(15.22)
j =
ih
2m
(

). (15.23)
Note that the reality of the function V was crucial for this result. If we used a complex
potential energy this continuity equation would not arise (see problems).
We can now use our previous experience with continuity equations to derive a conser-
vation law. Recall that a continuity equation such as (15.21) implies that the time rate
of change of the volume integral of over a volume R will be controlled by the ux of j
through the boundary S of R. Thus,
N(t) =
Z
R

(r, t)(r, t) d
3
x (15.24)
satises
dN(t)
dt
=
Z
S
j ndS, (15.25)
provided, of course, that satises the Schrodinger equation (15.14). In particular, with
boundary conditions chosen so that the ux of j through S vanishes, the probability for
nding the particle in the region R will be time-independent. One says that probability
is conserved.
This conservation law allows us to normalize the solutions to the Schrodinger equation.
Recall that (r, t) d
3
x is the probability that the particle is located in an innitesimal
neighborhood d
3
x of r at time t. Suppose that the particle is restricted to a region R of
space (which may in fact be all of space). The total probability for nding the particle
anywhere in R at any given time should be unity. Thus we should demand that at any
time t Z
R

(r, t)(r, t) d
3
x = 1. (15.26)
138
One says that the wave function is normalized (to unity); this normalization is crucial
for the physical interpretation of the wave function in terms of probabilities. In particular,
the initial (say, t = 0) wave function should be normalized:
Z
R

(r, 0)(r, 0) d
3
x = 1. (15.27)
In fact, it is enough specify initial/boundary conditions such that (15.27) holds and the
wave function is guaranteed to be normalized for all time if it satises the Schrodinger
equation. Indeed, if (15.27) is satised, then with boundary conditions chosen such that
the ux of j through the boundary of R vanishes, (15.25) guarantees that (15.26) is satised
(exercise). This result is quite important since the solutions to the Schrodinger equation
are uniquely determined by the initial (normalized) wave function. If the wave function at
later times were not normalized, then the probability interpretation of quantum mechanics
would not work.
15.5 Boundary Conditions, Particle in a Box
Our argument that took us from the Schrodinger equation to the continuity equation
and to the conservation of probability relied upon using appropriate boundary conditions.
Appropriate boundary conditions for the normalization are such that the ux of j through
the boundary of the region of interest should vanish. If this region is all of space, this is
accomplished by using solutions to the Schrodinger equation such that 0 (at a fast
enough rate) as r . Physically, this corresponds to requiring that the particle never
escapes to innity (at any nite time).
It is often physically appropriate to limit the spatial domain of the particle. A common
model system used in quantum mechanics is a particle in a box. This is a model of a
particle which is conned to some nite region in space. For example, a spherical box
would be the points r < a, a = constant, and we would demand that = 0 at r a.
This means that the particle is never outside the spherical box. Let us explore a simplied
model of a particle in a box in a little more detail.
We again restrict our attention to one spatial dimension for simplicity. We consider
a free particle moving in a box in which L < x < L. We look for a solution of the
free particle Schrodinger equation that is non-zero in the box, but is zero outside the box.
Since the zero function always satises the Schrodinger equation, we have already solved
the Schrodinger equation outside the box (exercise). We will restrict our attention to solu-
tions which continuously join this exterior solution, i.e., the solutions must continuously
vanish at the walls of the box:
(L, t) = (L, t) = 0. (15.28)
139
For example, a simple function which vanishes at the boundaries x = L is
f(x) = N sin

n
L
x

, (15.29)
where N is a constant (determined by normalization) and n is an integer. We require
the particle to be somewhere, so it wont do to let = 0 everywhere. Thus we restrict
attention to n 6= 0. Let us choose this sine function to represent the initital wave function
for the particle in the box:
(x, 0) =

N sin

n
L
x

, if L x L;
0, if |x| L.
(15.30)
The constant N is determined by normalization:
Z

||
2
dx = 1.
As an exercise you should do this integral in and show that, up to a phase factor of the
form e
i
, with real,
N =
r
1
L
. (15.31)
We suppose that (15.30) is the initial wave function. To nd the solution at time t with
this initial condition, we can use Fourier analysis, but let us take the following shortcut.
We already have a very simple class of solutions obtained using the separation of variables
technique (see (15.7)). The solutions shown there do not satisfy the boundary conditions
(15.28) because the function X(x) shown there does not satisfy the boundary conditions.
However both the Schrodinger equation as well as the ordinary dierential equations sat-
ised by X and T are linear so we can build new solutions by taking linear combinations.
And it is easy to take appropriate linear combinations to get solutions which do satisfy
(15.28) (exercise). You can easily check (exercise) that with
(x, t) = X(x)T(t), (15.32)
where
X(x) =
r
1
L
sin kx, T(t) = e
i(k)t
, (15.33)
and
k =
n
L
, (k) =
hk
2
2m
(15.34)
we get a solution to the free particle Schrodinger equation that is always normalized, sat-
ises the boundary conditions (L, t) = 0, and agrees with the initial condition (15.30).
Physically, the wave function (15.32)(15.34) represents a particle with energy
E
n
=
h
2

2
2mL
2
n
2
,
140
moving in the region L x L. ( Moving is a bit misleading because the probability
distribution for position (and momentum, too) does not depend upon time (exercise). To
see time-dependent phenomena one needs to consider superpositions of wave functions with
dierent energies.)
141